1. The new journey of an old concept

Accountability-talk is omnipresent in contemporary law and politics.Footnote ² In spite of its rather high currency, though, accountability does not partake in the select group of first-order political ideals: democracy, human rights, constitutionalism and rule of law have all been historically uttered in much more vocal tones and still remain at the forefront of public demands for legitimate authority. Accountability, in turn, has stayed at the backdoor of our legal or political vocabulary and operates at a lower waveband. It has not exactly been an incendiary enough flag to lead people to the streets, but it still incites reformist initiatives. It somewhat assists and qualifies each of those stand-alone ideals to pursue their respective ends.Footnote ³ Rather than radiating a comprehensive legal or political vision, accountability supplies a power-constraining toolbox that allows for a variety of permutations. Each permutation will hinge on the role, place and weight of the entity or actor to be held accountable.

This ability to serve various masters makes accountability all the more intriguing and chameleonic. To hold a powerful individual or an institution accountable is generally presumed to be a good thing. Accountability is announced, in other words, as a praiseworthy goal to be pursued by law and politics no matter where it takes place, be it locally or regionally, nationally or internationally. It would protect the account-holder and, in several ways, perhaps counter-intuitively, it may benefit the accountee as well.

This rather commonplace story, though, tells very little about the concrete configurations of accountability, the specific values and ends, if any, it is supposed to attain, let alone the exact settings in which it should apply or the functions it should fulfil. That resounding catchword, as it turns out, offers neither a straightforward definition nor a clear portrait of its own potential institutional translations. Therefore, it would be methodologically doubtful, if not futile, to start off any investigation about accountability without, first, cleaning up the haphazard rhetoric and, then, openly constructing a firmer conceptual foundation.Footnote ⁴

The concept of accountability is Janus-faced. Its duality is expressed in three distinct ways that are complexly intertwined with each other. The concept can: (i) display a descriptive or normative character; (ii) capture a political or extra-political relationship; and (iii) be shaped by legal or extra-legal properties. In order to get a satisfactory grip on that concept, one should put these three parallel dimensions into an adequate and coherent context. Should any of these three dimensions be left out of the picture, the explanatory capacity or normative appeal of the concept will significantly recede. Should they get inedvertently conflated and their connections obscured, the application of the concept becomes equivocal.

With regards to the first dimension, accountability can be conceived both as a given and as a construct, both as an inevitable ‘social fact’ and as a ‘purposeful enterprise’.Footnote ⁵ Mostly, however, it seems to convey nothing but noble intentions in the ordinary political lexicon. That is to say, its normative side usually overshadows its purely descriptive one. Converted into a ‘feel good’ term, it often ends up working as a verbal weapon for political strategists and rhetoricians, not as a tool for legal and institutional analysis.Footnote ⁶

Accountability, moreover, is not a solely political concept. A political agent, apart from making and implementing decisions, inevitably carries some burden of accountability for her actions. To some extent, moral agents, in several extra-political domains, also have to discharge a similar duty. The obligation to give an account to someone, somehow, for some particular act and at some point in time, and the expectation that this someone will react, positively or negatively, strongly or weakly, at some other point in time, cuts across the conventionalized borders between politics and other spheres of social life.

Finally, the concept may get more or less enmeshed with law. First, that might happen when the concept enters the province of institutionalized politics, a province constituted and governed by legal rules and procedures. Second, that might also happen when one confronts social relations that, despite not being purely political, have been densely legalized.

This article clarifies these manifold connections and delineates a framework to inform the analysis of the concept of accountability. To be sure, there is no single way to lay down such theoretical backdrop, no self-evident typology or uncontroversial boundary that the concept of accountability encapsulates. Every attempt of conceptualization passes over diverse aspects of the phenomenon it seeks to make sense of.

The varied literature thereon oscillates between more restricted and more expansive concepts of accountability, without incurring, necessarily, analytical inconsistency.Footnote ⁷ The concept of accountability stipulated in this article is a minimalist one, that is, comprehensive enough to grasp different sorts of power relations. I do not claim this is the one and only correct definition of the term, but it nonetheless illuminates accountability facets of a broader spectrum of transnational institutions than a maximalist concept would otherwise do.Footnote ⁸ The more minimalist the definition, the larger its denotative range naturally becomes. At the transnational level, to avoid ‘overloading the definition of accountability’Footnote ⁹ may be cognitively productive to perceive more nuanced interactions that shape accountability relationships, no matter how attractive they may be from the normative point of view.

According to such a minimalist concept, accountability directs itself to the phenomenon of power. It hinges upon the divide between power-holders and subjects to power. It establishes a peculiar and contingent sort of relationship between both sides (by way of converting the latter into an ‘account-holder’).

Basically, this article elucidates what accountable and unaccountable powers essentially are (and the gradual variations between both poles), and why the latter is deemed a rather alarming entity, while the former is quite vehemently cherished. Defining the concept and distinguishing its descriptive and normative, political and extra-political, legal and extra-legal aspects are major analytical steps undertaken by the accountability literature.Footnote ¹⁰ This article aims to make sense of these various facets in a more systematic fashion. Let me briefly explain how.

The next section (Section 2) characterizes political accountability as opposed to its extra-political forms and advances the basic structure of any accountability relationship. It answers the following two questions: what does it mean to be accountable? What does it mean, in particular, to be politically accountable? Although political and extra-political forms of accountability share a common structure, the article is mainly concerned with the ways in which political authorities, in particular transnational ones, may be held to account.

Section 3 points out the fundamental co-ordinates of political accountability, that is, the descriptive angles that help us to decode and to draw an analytical map of any power relationship that display some accountability nuance. Among other things, this section spells out how law may be present or absent in political accountability, rejecting the ambiguous dichotomy that opposes legal and political accountability.

Finally, Section 4 sets out the key functions political accountability has been expected to play. If one plans to go beyond description, as the political discourse and the literature in law and international relations have actually gone, in the name of what purposes should one couch the demands for accountability? Answering such question calls attention to the normative side of the concept of accountability and, more importantly, the occasional tensions that might arise.

From Section 2 to Section 4, thus, we can see a gradual knitting of distinctions that culminate, hopefully, in a more detailed picture of the concept of accountability. There is a thread that holds these sections together: from the concept of accountability and its descriptive co-ordinates, we realize that accountability is a compounded link between a power-holder and someone else; still, it is only by the elucidation of the functions that we begin to understand when and why accountability is a desirable attribute in the first place and why it should be fostered.

This article, thus, purports to capture the ‘complex dimensionality’ of accountability.Footnote ¹¹ The distinction between the dimensions, co-ordinates and functions, and a clear picture of how they relate, are primary steps for adequate understanding and well thought out institutional reforms.

2. Political and extra-political accountability: shifting and contested boundaries

Accountability is not, by definition, restricted to politics, no matter how politics is conceived.Footnote ¹² It transcends the domain of authoritative collective action and pervades social relations more widely and deeply.Footnote ¹³ One should be attentive to extra-political forms of accountability in order to understand what accountability shares with the political form and what is distinctive of the latter. Rather than a pedantic digression, this inquiry helps us both to single out, at least approximately, the elementary particle of the concept and to discern the concept itself from its uniquely political manifestation.Footnote ¹⁴

There are several types of extra-political accountability.Footnote ¹⁵ Cultural or social accountability is probably the most difficult to perceive and diagnose. When feminist activists, for example, complain about the ‘failure of the culture to hold men accountable’,Footnote ¹⁶ they do not necessarily ask for legal punishment or for the use of coercion to right the wrongs allegedly committed by men. Their protest runs deeper. It concerns the more diffuse and subtle ways through which a reputedly gender-biased culture curtails the autonomy of women to lead their own lives. The demand for cultural accountability, in this sense, intends to unveil and to challenge the sources of social normativity as well as the informal ways through which standards of behaviour are furtively enacted and enforced in a potentially oppressive fashion. Accountability mechanisms, thus, would be constitutive of a more horizontal social fabric.Footnote ¹⁷

Accountability relations also emerge in a variety of other extra-political settings.Footnote ¹⁸ Professionals, for example, may be called to account in the light of how they exercise their expertise vis-à-vis their peers and laypeople.Footnote ¹⁹ Lawyers, engineers or medical doctors take technical and learned decisions the appropriateness of which their respective clients or patients have limited capacity to judge. Still, this would not necessarily prevent the latter from having a right to demand, from the former, a justification for the choices that were made as well as from taking a stand on such choices’ acceptableness and accuracy. Their reputation towards their peers, moreover, may be a decisive factor for their professional success and self-satisfaction.

Corporations may be called to account for how their market actions affect the lives of their clients, employees or society at large; for what their products stand for; for how these products are made and advertised; for which patterns of consumption are encouraged; for how their employees are treated and how their facilities are constructed or, more broadly, for where their money is spent and investments are addressed. The answers to all these questions impact on different social groups that may wish or reasonably deserve to have an input on decisions that, one way or another, also concern them.Footnote ²⁰

Cultural, professional or corporate accountability share something in common. In all these cases, there is a split between an agent who takes decisions and another who bears the impact of or has some stake in these decisions. There is, thus, a division of labour between two poles. Accountability is a quality that may or may not permeate the relationship between both groups. It exists where the decision-maker has the obligation or is factually impelled to account, and where the subjects to the decisions are entitled or factually able to demand an account for the actions or inactions of the decision-maker.Footnote ²¹

Accordingly, accountability monitors or even compensates for that division of labour by constraining, to varying degrees, the autonomy of the decision-maker. Hence, the behaviour of accountable decision-makers is far from unbound. Exogenous considerations will need to be factored into their decisions, which interfere with the decisions themselves. For sure, the force and the kinds of constraints will vary, and the extent to which they can credibly curb one's actions will also differ across time and space.

Accountability revolves around a power-holder and a ‘significant other’.Footnote ²² The ‘other’ is only ‘significant’ as far as she has a credible claim on the power-holder, that is, the ability of demanding the latter to furnish an account for her conduct, of engendering, to use Bovens’ phrase, a ‘reflective discursive encounter’.Footnote ²³ It is reflective since it lets the agent turn inward and find a justification for her acts; it is discursive because the agent also needs to turn outward and express this justification publicly and intelligibly in order to engage in a back-and-forth conversation; and, furthermore, it instantiates a non-arbitrary encounter inasmuch as there is some sort of link binding the two subjects. This encounter, to sum up, amounts to ‘an ongoing process of account-giving and account-taking’.Footnote ²⁴

However large the diversity of accountability relationships can be, and despite their particularities, there is a core analytical structure that imbues all the examples above and below. Such structure can be enclosed by a set of rudimentary descriptive questions: Who accounts to whom? For what and on the basis of which standards? How and when? Under pain of what consequences?Footnote ²⁵

The first question brings forward the subjects of an accountability relation, or who the accountee and the account-holder are.Footnote ²⁶ The second specifies the object and standards of accountability, that is, the acts of power that are exposed to such constraint (after all, power-holders may not be accountable for acts that are not related to the exercise of such power), and also the benchmark of judgment that informs such account. The third settles the procedure and timing of an accountability relation, or the way it is done and the moment it takes place. The fourth, finally, prescribes the consequences that will ensue.

We can also turn this descriptive question into an explanatory one: Why does A account to B for the K acts, on the basis of X standards, through Y procedure and at time Z? The answer would have to spell out what is the causal story that binds A to B and its respective connection with K, X, Y and Z. Such answer would give us a thorough picture of how a definite accountability relationship is conformed. They still do not suffice, however, to grasp the normative dimension that a plea for accountability entails. Bringing a critical grip to the question, then, one could ultimately ask: who should be accountable to whom for what, on the basis of which standards, how and when?

The aforementioned descriptive, explanatory and prescriptive variants of this structural question are the key analytical prisms to understand the disputes and arguments about accountability. Accordingly, any attempt to classify an accountability relationship into this or that category will revolve around this fundamental formal pattern. Whether the accountability relationship is political or extra-political will depend on the contingent question of how we feed each of those components – the subjects, the object, the standards and so forth.Footnote ²⁷

The uniqueness of political accountability relates to the sort of power in play, namely, political power, to the subjects to which the exercise of that power is directed – the members of a political community – and to the issue that is under consideration – the general interest, the common good or the like.Footnote ²⁸ Decisions taken under such political authority are binding on and made in the name of the polity. Collective decisions based on the political pedigree may generate a duty of obedience whose enforceability does not depend upon the individual endorsement of the content of every single choice. The effectiveness of such collective decisions, as it happens with politics, presupposes a sense of trust and membership.Footnote ²⁹-Footnote ³⁰ However blurred the boundaries between political and extra-political types of accountability might be in the edges, the distinction between these two types is worth drawing. Yet this does not mean that their relational structure is substantially different.Footnote ³¹ Quite the opposite is the case: as far as the structure is concerned, political and extra-political accountability share much in common.

In sum, a power-holder may be deemed accountable when her power is plausibly constrained by the expectations, instructions or stakes of one or several external agents.Footnote ³² She is thus compelled to furnish an account for what she is doing or not doing. Such rudimentary definition, for sure, is still far from sufficient to fully equip the inquiry that this article intends to undertake. For instance, one could argue that it is too vacuous to come to terms with the contemporary debates about accountability deficits and gaps, overloads and excesses in either national or international levels.Footnote ³³ This sketch must not be the entire notion of accountability these debates have in mind. After all, there is hardly an absolutely unaccountable power, political or otherwise, and some forms of accountability may be spurious rather than admirable.Footnote ³⁴ Their protest calls for something more specific and normatively loaded, not for whatever sort of accountability one happens to discern.

In fact, the sheer existence of accountability is no reason for celebration. Politics is often embroiled in a string of more or less onerous accountability relationships. Behind the veneer of public rhetoric, the actual discussion has rarely been about whether a certain agent is accountable.Footnote ³⁵ The critical question has rather been about whether one is accountable to the right constituency, in the right way, at the right time, for the right reasons or standards. Thus, in order to get some traction, one needs to delve deeper into these thorny normative elements. One does not contribute to this debate without openly recognizing these normative underpinnings. Bracketing them is not a sustainable option because a value-free standpoint is simply not available.

3. The co-ordinates of political accountability

The field of application of political accountability is vast. It comprises diverse sorts of political relationships, which are assimilated by an array of typologies.Footnote ³⁶ The regimes of political accountability resort to a number of devices that inform the interaction between accountees and account-holders. Here I introduce some of the chief categories that help to particularize these several possible arrangements.Footnote ³⁷

The structural question put forward in the previous section specifies a way to envisage the co-ordinates of political accountability (or, in fact, of any power relationship): the subjects, the object, the benchmark of judgment, the procedure, the timing and the consequences that bear upon such a relationship. Here, I spell out the co-ordinates that, combined in different ways and doses, allow accountability arrangements to fulfil different functions (as the next section will show). What follows comprises eight co-ordinates of accountability. This list is surely not exhaustive. The goal, however, is to gain systematicity and to group the properties of accountability that are still fragmented in the specialized literature.

Every accountability relationship has a distinct density. Depending on how its co-ordinates are configured, it will vary across thicker and thinner ends of a spectrum of possibilities. Thicker and thinner modes, as we shall see, imply not only that the accountees will feel more or less constrained by (or tied to) the account-holder, but also that it will be enmeshed in more or less complex, institutionalized and formal sorts of connection.

The first co-ordinate through which that density can be observed purports to capture the level of formality of an accountability relationship.Footnote ³⁸ This is a leading co-ordinate that pervades and shapes all others. Any power relationship will be more or less formal and legalized. This means that general and previously enacted legal rules may regulate each of the co-ordinates listed below, ascribe an official status to their subjects and specify their respective authority and competence. Accountability relationships may benefit from the qualities of the rule of law in order to become more stable and predictable.Footnote ³⁹

This legal angle invites a conceptual caveat. It would be erroneous, of course, to equate the distinction between political and extra-political accountability with the presence or absence of law (or adjudication). Law is only one of the several possible institutional conveyors that can promote both political and various sorts of extra-political accountability. It brings to accountability relationships a measure of rule-like formality and the accompanying apparatus for legal enforcement. Legal norms and institutions are commonly used for promoting corporate, professional or parental accountability, as well as for fostering political accountability. Law, however, does neither exhaust their potential nor capture informal or less rule-bound accountability tools that may also operate within each of these accountability relationships, including the political ones. Perceptible traces of accountability still subsist both in ‘law-free zones of privacy and association’Footnote ⁴⁰ or in law-free zones of political relationships, which should not be ignored. This explains why the categorical opposition between political and legal accountability, proposed by some typologies, may be misleading part of the time.Footnote ⁴¹

The second co-ordinate resorts to a traditional spatial metaphor. It regards how the power relationship between two agents materializes along horizontal or vertical lines. The vertical angle is more typical of accountability discourses. Substantially, vertical accountability is characterized by some asymmetry of power between accountees and account-holders. It can run downstream when the account-holder delegates power to the subject that will then be held accountable (in a ‘principal-agent’ or ‘truster-trustee’ fashion) or it can run upstream when the accountee is held accountable by those who are bound by her decisions or somehow bear their impact. To use a fashionable dichotomy, a vertical accountability relationship can run either top down or bottom up. The same actor, moreover, may be accountable both ways – from above and from below.Footnote ⁴²

Apart from this vertical axis, there can also be accountability relationships that take place along a horizontal (or gradually less vertical) line.Footnote ⁴³ This may occur between two agents that are tied neither by a clear relation of ‘command-and-control’ authority nor by an unmitigated duty of obedience. The constraint, in this case, is more delicate and stems rather from a co-operative commitment in light of mutual dependence. Scrutinizing it further, a horizontal accountability relationship could be unidirectional if, despite the horizontality, only one of the subjects is truly held accountable to the other, or bidirectional, if there is a reciprocal constraint to account for the decisions that either subject takes. ‘Checks and balances’ and its several internal tools for mutual control and co-operation are classical instance of this horizontal bidirectional type, where a single agent is both an accountee and an account-holder at one and the same time.Footnote ⁴⁴

The third co-ordinate highlights a relevant nuance of accountability relationships in multi-member institutions. It shows whether it is the institution, organically considered, or each of its members, who is called to account. The process of holding a power-wielder accountable can work in the wholesale and in the retail or, to use another common dichotomy, be centripetal and centrifugal. In the former case, the entire collegiate (like a court, a regulatory commission or any deliberative body), rather than its members, may be assessed and held responsible for its decisions. In the latter case, on the contrary, the institution evades a collective responsibility, but each member is assessed by the way she has contributed to the collective performance.Footnote ⁴⁵

An elected parliament, in a constitutional system that also adopts judicial review of legislation, simultaneously provides an example of both: on the one hand, the parliament is subject to atomized accountability by the constituents, who are able to judge, reward and punish, by means of election or non-election, each representative, but not the institution itself;Footnote ⁴⁶ on the other hand, the parliament is subject to organic accountability by the court that asseses the constitutionality of legislation. In such a system, the parliament is held to account both as ‘many’ and as ‘one’, both in its plurality and in its supra-individual unity.

An accountability relationship may also be marked by demands of expertise. Through this fourth prism, one should realize whether the subjects possess special knowledge, that is, whether they are technocrats and hence decide on the basis of reputedly objective, impartial and universalizable premises.Footnote ⁴⁷ Here, four pictures may conceivably emerge between accountees and account-holders: (i) both are experts, which would lead to a strictly technocratic accountability relationship;Footnote ⁴⁸ (ii) both are lay and nontechnical agents;Footnote ⁴⁹ (iii) the accountee is an expert whereas the account-holder is not;Footnote ⁵⁰ and (iv) the accountee is lay whereas the account-holder is an expert, shaping a sort of guardianship regime.Footnote ⁵¹

Power-holders are usually expected to discharge a rather specific role. The performance of this task might be accompanied by substantive standards and expectations to appraise the output that is delivered (which might be related to expertise or not). The fifth co-ordinate, thus, reveals whether there are explicit criteria through which the performance of power-holders will be evaluated. The specificity of standards may vary and hence provide, to the power-holder, more or less precise guidelines on how her performance will be rated. The existence of public standards (or the lack thereof) will qualify, for example, both the exercises of voting and reason-giving, mentioned below.

The stringency of standards shapes diverse kinds of accountability relationships. Whatever their sources are, the more standards constrict the behaviour of the power-holder, the less room for the exercise of discretion will remain. The openness of standards turns an accountability relationship into one between a ‘truster’ and a ‘trustee’. At the other end of a continuum of stringency, the relationship might be reduced to one between a ‘rule-setter’ and a ‘rule-follower’, where discretion itself might disappear (and, with it, the very idea of accountability).Footnote ⁵² If accountability is not a matter of strict compliance with rules and standards, so as to render accountees ‘wholly subservient’Footnote ⁵³ to account-holders, and if it should rather encourage a mindful exercise of judgment, some degree of agency for the power-holder has to be preserved. Wherever this threshold might be located, one might expect to see variations of degree along this continuum. As Philp suggests, at one pole there will be a ‘compliance-based system’, structured around a set of incentives for conformity and threats against non-compliance; at the opposite pole, there will an ‘integrity-based system’, which prioritizes agency over rule-following and hinges upon trust.Footnote ⁵⁴

The sixth co-ordinate exposes the modus operandi of an accountability relationship, or, in other words, how proceduralized it is. Several methods of participation, enquiry and contestation may sew a web of constraints on decision-making. The way in which the accountee is appointed, appraised or even removed from office by the account-holder may be governed by more or less consolidated practices and conventions. Voting and reason-giving are two stereotypical mechanisms through which these procedures are designed. By voting (or simply nominating), the account-holder can place someone in or, without further justification, remove someone from a position of power. By having the occasion to reason and interrogate (through a ‘notice and comment’ mechanism, for example), the account-holder can also air disagreements that put pressure on the accountee, triggering a sort of deliberation between them both.Footnote ⁵⁵ The power-holder, in turn, may have the obligation to articulate a public justification for each decision she takes.Footnote ⁵⁶ The accountability-promoting quality of reason-giving, therefore, would be able to operate in two directions: for the account-holder, as a right both to give and to receive reasons; for the accountee, as a pressing burden. To put it differently, while the former may constrain through the opportunity to reason, the latter is constrained by the duty to reason.

A variety of balances between transparency and confidentiality will ultimately delineate how the procedures of voting, arguing and commenting by the account-holder, or the onus of reason-giving by the accountee, set the stage for an altogether plausible accountability relationship.Footnote ⁵⁷ Rather than accountability itself, the transparency of the procedures and decisions taken by the accountee constitutes the informational pre-condition for such a relationship to be set in place. On the whole, transparent authority is not, in and of itself, accountable. Accountable authority, though, cannot but be a minimally transparent one.Footnote ⁵⁸ Total opacity and accountability do not match.

Accountability interactions may also be put into a temporal perspective.Footnote ⁵⁹ This seventh prism illuminates the exact moment the power-holder expects a reaction from the account-holder, or the moment the latter has the opportunity to respond to the former. Two main temporal co-ordinates help discerning the possible variations. Firstly, the accounting may be ex post or ex ante, that is, it may take place after the decision by the accountee is made and implemented, consummating more or less concrete effects, or it may come somewhat earlier.Footnote ⁶⁰ Mechanisms of preventive control illustrate the latter whereas the posterior ascription of responsibility exemplifies the former. A combination between the two possibilities may jointly form a single accountability relationship. Occasionally, when the power-holder anticipates the potential reactions of the account-holder and acts accordingly, the distinction itself may partially lose its grip. The distinction between ex ante and ex post, thus, captures an ambivalence in the phenomenon of ‘being accountable’: to ‘perceive yourself accountable’ (and acting accordingly) and to ‘be held accountable’ (and suffering the consequences of your previous acts) are not the same thing. Both may indeed coalesce part of, or even most of the time, but it is crucial to note when they do not.

Secondly, both ex ante and ex post reactions could also be qualified as immediate and prompt or as diffuse and gradual. Such temporal angle enables one to insert accountability relationships into a short or long-term horizon. Thirdly, these interactions across time can be characterized as a process and hence considered as following a set of phases. Roughly, this process would comprise, as Keohane and Grant suggest, three phases: first, the standard-setting; second, the process of informing, justifying and assessing; lastly, the sanction.Footnote ⁶¹ This sequence of phases, as an ‘endless loop’,Footnote ⁶² can permanently restart and configure iterative ‘cycles of accountability’.Footnote ⁶³

Finally, the eighth co-ordinate refers to the weight of the consequences that ensue from the exercise of power. The amount of power bestowed upon account-holders to sanction accountees varies. Sanction is a wide concept and comprehends mechanisms with greater or lesser teeth. Some authors have equated the power to sanction to the power to remove someone from office (a typical example would be the power of voters to periodically elect, re-elect or ‘deselect’ their representatives). This automatic equation, though, neglects the more subtle ways through which power may effectively be sanctioned and held accountable.

The power to remove the accountee is probably the most drastic but by no means the only instantiation of sanction there can be. Sanctions that fall short of sheer removal may comprise forms of public criticism and exposure, shaming and stigmatizing, all of which can jeopardize the exercise and the very viability of legitimate power.Footnote ⁶⁴ The shrinkage of reputation, for example, might be disastrous for decision-making bodies that operate within a terrain of soft power.Footnote ⁶⁵

Hard and soft or direct and indirect are apposite dichotomies to enclose the variety of sanctions there may be. It is undeniably important, for sure, to keep in mind that ‘inconsequential accountability is no accountability at all’.Footnote ⁶⁶ Embracing this forceful maxim, however, does not imply reducing ‘consequential’ to the most severe kind of sanction. Whether a soft or indirect sanction creates efficient and credible enough constraints on the power-holder is a different, if relevant, empirical question, the implications of which should be addressed separately, according to each context.Footnote ⁶⁷ Some of them will do, while others will certainly not. Generalizations in such a domain would be prone to failure.Footnote ⁶⁸ At any rate, the extent to which sanction is a ‘definitional component’ of an accountability relationship will depend on how one defines sanction itself.Footnote ⁶⁹

These are, in sum, the minimal properties of an accountability relationship. With such co-ordinates at hand, one should be able not only to answer the set of descriptive questions advanced by the previous section in a more thorough and meaningful way, but also to join the normative debate from a more comprehensive starting point. Without such an analytical map, the observer may not go too far in portraying and criticizing so complicated arrangements. Institutional architects and builders have a momentous task: first, the one of identifying each single piece in an institutional paraphernalia; second, to understand how each of them is causally implicated; finally, to discover what calibration of each co-ordinate would favour each of the normative functions that accountability arrangements are expected to play (as explained in the next section, the constitutional, democratic, epistemic and populist).

The following table pinpoints these elementary units:

One can certainly play with these abstract ingredients and speculate about which recipes would engender the thickest and the thinnest relationships. Authors widely diverge on finding the exact thresholds, and occasionally deny the title of accountability to some of its thinnest instances.Footnote ⁷⁰ Here, I do not intend to engage in this sort of definitional line-drawing. The co-ordinates offer a broad picture of how power relations are configured and how power-holders may be constrained and held accountable.Footnote ⁷¹

When pressed against complex concrete arrangements, these mostly dichotomic categories may fall apart somewhere at the edges. Abandoning these categories, however, would give away a useful source of insight, an angle without which, I believe, important things would otherwise remain unsaid and sound hardly plausible. A reliable description necessarily precedes competent prescription, and that is the cognitive service provided by the co-ordinates to the functions. The next section sets out the main functions one can derive from a descriptive portrait.

4. The functions (and dysfunctions) of political accountability

Institutions are not built in order to be accountable in the first place.Footnote ⁷² Some way or another, however, they are likely to be. If we needed a concept to help us minimally describe the features of a specific power-relation, the distinctions elaborated above would fairly do. However, in order to gain normative traction – or to know both who, how and why should account, and who, how and why should hold to account – one has to further engage with the values and purposes involved. That would presuppose, to start with, a conception of legitimate politics.Footnote ⁷³ Accountability conveys no more than a special link between a power-holder and someone else. The mere perception of such a link, in itself, leaves us in a normative vacuum.Footnote ⁷⁴ One can presume neither a positive nor a negative quality in it. There is no such thing as a ‘pure theory’ of accountability, as far as a value-laden prescription is concerned. That is, the appeal of the contemporary call for accountability does not stem from accountability tout court. Such call is not self-standing, but rather ancillary to an external ideal, be it explicitly articulated or not.Footnote ⁷⁵ Or so I shall argue.

The most plausible way to defend and justify accountability, in this sense, is instrumentalist. Instead of an end in itself, it is a means to an end.Footnote ⁷⁶ More precisely, it is a means to a series of dissimilar and usually conflicting ends that those external ideals articulate. There is no single self-evident end to be promoted. This discomforting feature leads to an inevitably contested dispute.Footnote ⁷⁷ Which are, then, the expectations that are embedded in current accountability-talk? What are the institutional ills it is supposed to cure or the vices it is expected to prevent? When can accountability, to sum up, be seen as a good thing, as a positive qualifier of power relationships?

This is an intricate theoretical problem, as controversial as political philosophy can get. To take a firm stand on such dispute is beyond the scope of this article. Nevertheless, one cannot entirely dodge the question if one wants to somehow participate in such debate. If anything, it is crucial to clearly identify the acceptable normative goals of accountability and be prepared to recognize that, depending on the context, these goals may be mutually inconsistent or reinforcing. It is up to the context-oriented institutional designer to strike the suitable balances and value judgments. Tensions can hardly be solved, but a conscious designer has to manage them.Footnote ⁷⁸ Analytical transparency is a pre-condition for meaningful dialogue on the matter.

Accountability scholarship roughly puts forward four pivotal normative rationales to back up its claims. Accountability devices would orient themselves (i) towards limiting power and inhibiting abuses; (ii) towards recognizing, listening and responding to the plurality of voices of the account-holders – those who are deemed to have legitimate stakes on the matter; (iii) towards building institutional capacity – a particular craft for taking substantively good decisions; or, finally, (iv) towards fostering allegiance and obedience from the account-holders. Each rationale has a different story of legitimacy to tell, a distinct reason as to why accountability is a desirable attribute.Footnote ⁷⁹ The first rationale is defensive, the second is emancipatory, the third is technical and the fourth is strategic. I will call them, respectively, the constitutional, the democratic, the epistemic and the populist ambitions of accountability.Footnote ⁸⁰ Such polemical terms evoke concepts embroiled in complicated tensions and overlapping zones, not to mention their diverse yet intertwined historical traditions, which I will not directly address. I resort to them as terms of art.

Let me elaborate a bit more on these four aspects of legitimacy. The constitutional aspect reverberates a conventional way to justify accountability. It basically seeks ‘to keep power from running wild’,Footnote ⁸¹ to moderate and counterbalance its weight through a set of procedural techniques and substantive standards. Under such perspective, legitimate power cannot be raw and naked power. It should be restricted, so that the arguably constant danger of abuse (that is, of overstepping the commonly regarded limit) gets domesticated. This political maxim is meant to serve a clear-cut goal: the protection of individual autonomy. Accountability, here, enables account-holders to oversee their accountees, to ascribe responsibilities to them and to somehow punish or reward them, if necessary. The point of constraining authority, thus, is to block arbitrariness, to retain power under check.Footnote ⁸²

The democratic variant of accountability is concerned with something else. Regardless of limits or checks, it is meant to give any member of a political community who might have been affected or otherwise influenced by the exercise of authority, some fair leverage in collective decision-making. The usually rhetorical appeal of ‘the people’ as the one and only source of legitimate power is, in practice, translated into more or less convoluted chains of power delegation or ‘transmission-belts’.Footnote ⁸³ If authority is accountable in that way, one is supposed to trace back and ultimately find some incarnation of ‘the people’ in the decisions that ask for collective compliance. Procedurally, this means that some channel must carry the plurality of voices into the final decision-making table, however distant that table might be. These voices, to invoke three other mainstream terms, are ‘included’ or ‘represented’, and somehow ‘participate’ in that table.Footnote ⁸⁴ The farther away this delegation string goes, nevertheless, the quieter the people's voice and hence the less credible the democratic vestige becomes.Footnote ⁸⁵ How far-off that string can be stretched for convincingly granting the democratic pedigree is a disputed issue.

Accountability arrangements, however, do not only interpose safeguards between power-holders and those subject to power in order to prevent abuses (constitutional rationale). Neither do they solely seek to receive the inputs of ‘the people’ and somehow reflect and express their will, however loosely that will is conceived (democratic rationale). It is not only a transactional cost that hampers the efficiency of decision-making for the sake of other evenly important political goods. Accountability, as it has been argued, also carries a ‘promise of performance’Footnote ⁸⁶ or an ‘institutionalized capacity to learn’.Footnote ⁸⁷ It aspires to enable the power-holder to take appropriate decisions, to become an instrument for epistemically better choices. In that respect, accountability would improve the cognitive abilities of the institution and have a bearing on efficiency itself.Footnote ⁸⁸

Quite often, the concern with a putative accountability deficit is related neither to the risk of occasional power abuse nor to the lack of popular embedment. It is, rather, an apprehension about a possible malfunction, about the danger of power being used unwisely and unskilfully, even if, presumptively, with no arbitrary intentions, or arguably, with no elitist or exclusivist predispositions. In that case, accountability would not be dispensable no matter whether we could prove, hypothetically, that power-holders were the pure embodiment of reasonableness, self-restraint or popular conscience (that is, no matter how power-holders could fulfil our constitutional and democratic demands).Footnote ⁸⁹ This is a less visible aspect of an accountability relationship and must be elaborated with care.

Somehow, there is a banal connection insinuated here. After all, if, through echoing a prevalent normative premise of western political thought, we argue that arbitrary and obscurantist decisions are necessarily wrong, then accountability devices that manage to obstruct these sorts of decisional faults should presumably enhance competence. Such negative sense of the epistemic promise does not exhaust the present claim though.Footnote ⁹⁰ Otherwise, the epistemic ambition would simply be redundant, that is, would be conflated with either the constitutional or the democratic ones. Nonetheless, a well-crafted accountability relationship, some will argue, has a competence-based virtue for reasons that do not simply replicate the constitutional or democratic ambitions.

One could try to identify the singularity of the epistemic point by arguing that the pathologies of abuse and impermeability, which the constitutional and democratic dimensions respectively intend to rectify, are formal, whereas the epistemic point is substantive. The same political decision, thus, could be abusive or impermeable from the formal point of view, but still correct from a substantive perspective, or vice-versa.Footnote ⁹¹ But to classify the constitutional and democratic demands as merely formal, and to reserve the substantive pull for the epistemic facet, would neglect the intricate interdependence between form and substance that both ideals entail.Footnote ⁹² It would strain this formal/substantive dichotomy and yet offer an artificial answer to the question of which promise of accountability has a bearing upon the decisional output.Footnote ⁹³

The epistemic promise is actually more aspiring. It affirms that the performance of an institution is, to some extent, contingent on the sort of accountability arrangement that applies to it.Footnote ⁹⁴ The epistemic promise is not only concerned with avoiding abusive and impermeable decisions, even if we envisage them from a substantive point of view. It also aims to highlight how accountability devices may develop, depending on the context and decisional issue at stake, the aptitude or the proficiency of an institution to reach better decisions (according to some criterion of correctness). It purports to call our attention to the fact that, oftentimes, there is an intersection between being ‘accountable’ and being ‘competent’, a correlation between accountability devices and decisional accuracy.Footnote ⁹⁵ Among other things, then, choosing the adequate accountability regime would also presuppose ‘weighing the comparative competence or incompetence of alternative accountability arrangements’.Footnote ⁹⁶ An accountability device is hardly indifferent to that.

Lastly, accountability also has a populist aspect.Footnote ⁹⁷ Institutions tend to obtain better compliance with their decisions through mustering, with the aid of accountability mechanisms, a positive public perception. This property strengthens the reputation that an accountable institution conveys as well as such an institution's consequent ability to encourage actual obedience. It illuminates a plain causality: if an institution is able to ‘maintain the confidence of the public’,Footnote ⁹⁸ it has a greater propensity to be accepted. Accountability would help us construct, maintain and protect the trustworthiness and hence the social effectiveness of an institution.Footnote ⁹⁹-Footnote ¹⁰⁰

The co-ordinates and functions of accountability arrangements are thus closely implicated. Those descriptive variables can be blended, in each context, in order to perform the four legitimate functions of accountability. Among the infinite permutations of co-ordinates one could think of and experiment with, one can enumerate a few that are more frequently found in existing institutions and then discern their functional interrelation.

As for the democratic function, nowadays, one can hardly acknowledge that a political regime is democratic if it does not, at the very least, collect and convert individual votes into a quota of seats within a representative body through a reliable procedure (usually through an arithmetic formula that materializes a plausible conception of equality, as ‘one man, one vote’). Competitive elections, the act of delegation through which a principal-agent relationship between constituents and representatives is founded, are the first-best example of democratic accountability. Voting is the basic procedural mechanism to empower or disempower, to reward or sanction the representatives according to their performance as appraised by the constituents.

Second-best supplements, to be sure, can also resonate the democratic idea. Devices of participation, responsiveness, transparency and the like, despite engendering spheres of engagement and influence that fall short of decision-making power, establish horizontal constraints that can conceivably be traced back to the people. All these mechanisms, admittedly, tend to be primarily but not exclusively associated with democratic accountability, since institutional arrangements are versatile to pursue a large variety of ends. Elections, for example, are first and foremost perceived as a tool to empower the people. However, one should not underestimate their ability to limit governmental power, to trigger adequate legislation and policies and to induce the compliance of the constituents with the decisions of representatives (that is, to pursue the three other functions of accountability at the same time).

Constitutional function, in turn, are of two sorts: on the defensive and negative side, the procedure of horizontal checks and balances and the substantive standards that underlie the discourse of rights put, when in good working-order, limits on power by demarcating what power cannot do; on the constructive and discursive side, institutions in charge of deciding on the basis of principled and public reasons, like courts and other regulatory agencies, also flesh out the ambition of constitutional accountability.

As for the epistemic function, a variety of procedural mechanisms are regarded as appropriate to fulfil the specific task of an institution. When this task is somehow related to expertise, arrangements such as panels or committees of experts, dense mechanisms of peer review and review bodies, among others, may fit the bill.

Finally, the populist function of accountability does not exactly have a clear institutional configuration. Since its goal is to gain the allegiance and respect from the addressees of power, whatever arrangement can foster social legitimacy would meet its central demand.

To summarize, elections and a set of other voice-giving channels signal a democratic function; mechanisms of mutual oversight, substantive standards of decency and burdens of public justification play a self-styled constitutional role; devices of information-gathering and knowledge-producing would enhance an institution's epistemic capacities and, lastly, whatever mechanisms manage to improve the public perception of an institution would highlight a populist function.

These four dimensions, for sure, may crisscross each other from several angles when one uses them to observe concrete arrangements of accountability.Footnote ¹⁰¹ But the four ambitions do not always join forces. When their vectors clash or compete, some accommodation may prove indispensable.Footnote ¹⁰² Cases of divergence among the four functions tend to raise tricky normative dilemmas. If a point of equilibrium is not carved, the whole arrangement may become dysfunctional.Footnote ¹⁰³ Accountability becomes a laudable relational property to the extent that such balance is well struck.

Apart from the mismatch of different purposes, a quantitative sort of dysfunction might also arise when the amount of the accountability burden becomes counter-productive.Footnote ¹⁰⁴ Different sorts of deficits or excesses will emerge from the miscalculation of those functions. Processes of accountability deflation or inflation, thus, should be assessed with care. Inflation can be a welcome evolvement if the status quo characterizes a deficit, and vice-versa. Finding the optimal amount of accountability is a perpetual challenge of legal analysis and institutional design, which needs not to extrapolate the tipping point where accountability turns to be harmful rather than beneficial.Footnote ¹⁰⁵

Deciding how accountable an institution should be, thus, is not a straightforward and one-dimensional choice of an institutional designer. The appropriate accountability package is contingent on the ultimate purpose of an institution. The appropriate modular constructFootnote ¹⁰⁶ can be assembled only after such a purpose has been made sufficiently clear.Footnote ¹⁰⁷ Once fully developed, the modular construct has to delve into occasionally incongruous functions and to find out what devices better fit the particularities of this or that case. Conflicts of equally valuable or incommensurable functions demand the exercise of balancing or, to phrase it in economic jargon, imply trade-offs and cost-benefit assessments.Footnote ¹⁰⁸